[Telepathy] use adder to initiate correct audio stream

[Alexey Fisher]

Hallo all,

i still bothering me and you with the old performance problem. I know
that we can't imitate correct audio source without knowing what caps
audio codec will need. And we can't start the call without have working
stream.

Haw about this workaround:
- use "adder" (a gstreamer element) as switch between different audio
sources. Start with fake source: "audiotestsrc wave=silence !
audioresample ! audioconvert ! ..."
- after the call is initiated add new pad to adder
- initiate new source: for example pulsesrc
- block/remove pad with fakesource
- remove fake source.

I found one example in python:
http://stackoverflow.com/questions/3899666/adding-and-removing-audio-sources-to-from-gstreamer-pipeline-on-the-go